Systems and Methods to Process Electronic Images for Synthetic Image Generation

1. A system for composing a synthetic medical image, the system comprising: a data store for storing a plurality of medical images associated with data types; a processor; and a memory coupled to the processor and storing instructions that, when executed by the processor, cause the processor to perform operations including: receiving a request for a medical image having a data type; retrieving, from the plurality of medical images stored in the data store, a first medical image having the requested data type; retrieving, from the plurality of medical images stored in the data store, a second medical image without the requested data type; generating a synthetic medical image having the requested data type using the first medical image and the second medical image; and storing the synthetic medical image in association with the requested data type in the data store.

2. The system of claim 1 , wherein generating the synthetic medical image includes: arranging the first medical image and the second medical image side by side; at least one of: rotationally aligning the first medical image and the second medical image; or matching one or more image properties of the first medical image and the second medical image; and merging the first medical image and the second medical image to generate the synthetic medical image.

3. The system of claim 1 , the operations further including: receiving a semantic segmentation annotation for the first medical image associated with a region of the first medical image.

4. The system of claim 3 , wherein generating the synthetic medical image includes: identifying the region of the first medical image based on the semantic segmentation annotation; extracting the region from the first medical image; and injecting the region into the second medical image to generate the synthetic medical image.

5. The system of claim 1 , wherein the second medical image is selected randomly from the plurality of medical images for retrieval.

6. The system of claim 1 , wherein the second medical image is selected from the plurality of medical images for retrieval based on one or more features of the second medical image being complementary to one or more corresponding features of the first medical image.

7. The system of claim 1 , the operations further including: providing at least a portion of the plurality of medical images stored in the data store, including the synthetic medical image, and corresponding labels identifying the associated data types as a training dataset for input to a machine learning system to train the machine learning system.

8. The system of claim 1 , wherein the requested data type includes at least one of: an image modality, a target anatomical region, a target morphology, a presence or absence of a condition, or a presence or absence of a treatment effect.

9. The system of claim 8 , wherein the image modality includes digital pathology, magnetic resonance imaging (MRI), computed tomography (CT), X-ray, nuclear medicine imaging, or ultrasound.

10. The system of claim 8 , wherein at least one of the target morphology, the condition, or the treatment effect included in the requested data type is a rare presentation below a predetermined threshold of occurrence.

11. A method to compose a synthetic medical image, the method comprising: receiving a request for a medical image having a data type; retrieving, from a data store storing a plurality of medical images associated with data types, a first medical image having the requested data type; retrieving, from the data store, a second medical image without the requested data type; generating a synthetic medical image having the requested data type using the first medical image and the second medical image; and storing the synthetic medical image in association with the requested data type in the data store.

12. The method of claim 11 , wherein generating the synthetic medical image comprises: arranging the first medical image and the second medical image side by side; at least one of: rotationally aligning the first medical image and the second medical image; or matching one or more image properties of the first medical image and the second medical image; and merging the first medical image and the second medical image to generate the synthetic medical image.

13. The method of claim 11 , further comprising: receiving a semantic segmentation annotation for the first medical image associated with a region of the first medical image.

14. The method of claim 13 , wherein generating the synthetic medical image comprises: identifying the region of the first medical image based on the semantic segmentation annotation; extracting the region from the first medical image; and injecting the region into the second medical image to generate the synthetic medical image.

15. The method of claim 11 , wherein retrieving the second medical image comprises: randomly selecting the second medical image from the plurality of medical images stored in the data store.

16. The method of claim 11 , wherein retrieving the second medical image comprises: selecting the second medical image from the plurality of medical images stored in the data store based on one or more features of the second medical image being complementary to one or more corresponding features of the first medical image.

17. The method of claim 11 , further comprising: providing at least a portion of the plurality of the medical images stored in the data store, including the synthetic medical image, and corresponding labels identifying the associated data types as a training dataset for input to a machine learning system to train the machine learning system.

18. The method of claim 11 , wherein the requested data type includes at least one of: an image modality, a target anatomical region, a target morphology, a presence or absence of a condition, or a presence or absence of a treatment effect.

19. The method of claim 18 , wherein at least one of the target morphology, the condition, or the treatment effect included in the requested data type is a rare presentation below a predetermined threshold of occurrence.

20. A non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform operations for composing a synthetic medical image, the operations comprising: receiving a request for a medical image having a data type; retrieving, from a data store storing a plurality of medical images associated with data types, a first medical image having the requested data type; retrieving, from the data store, a second medical image without the requested data type; generating a synthetic medical image having the requested data type using the first medical image and the second medical image; and storing the synthetic medical image in association with the requested data type in the data store.